Hallo Joe,
Dienstag, 26. August 2008, meintest Du:
Questions node.
:
1) What is the cost-effective yet efficient way to connect this cluster
with IB?
JL Understand that cost-effective is not necessarily the highest
JL performance. This is where over-subscription comes in.
2) How many
Vincent,
I have always said that C++ is computational science's cold fusion:
lots of power in, but no net gain.
C is not much better. I once worked a young computational programmer
for almost a week to get him to prove to himself that a C source
program couldn't walk through a 2-d array
Michael H. Frese [EMAIL PROTECTED] writes:
C is not much better. I once worked a young computational programmer
for almost a week to get him to prove to himself that a C source
program couldn't walk through a 2-d array the hard way as fast as a
Fortran source program unless the stepping was
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Perry E. Metzger
Sent: Tuesday, August 26, 2008 8:06 AM
To: Michael H. Frese
Cc: Beowulf@beowulf.org
Subject: Re: [Beowulf] Stroustrup regarding multicore
Michael H. Frese [EMAIL PROTECTED] writes:
C is
I needed to parallelize a code already written as a serial code (with
domain decomposition in anticipation of the parallelization) by someone
who was teaching himself C++. The person put into the C++ everything
he was learning. From that experience I created my own personal list of
what to
Lux, James P [EMAIL PROTECTED] writes:
The example you give is going to be compiler dependent.
Even if it were, I defy you to find a c compiler that does
differently. As it is, this is not in fact compiler dependent.
There's no requirement in KR
There is, but in any case, KR is not the
Perry E. Metzger [EMAIL PROTECTED] writes:
I have a copy of the C99 document and it is indeed required that the
locations be consecutive (though there can of course be padding for
alignment purposes if you have an array of structures).
If you wish for me to quote chapter and verse from the
Lux, James P [EMAIL PROTECTED] writes:
No. You are actually given guarantees about memory layout. They're not
phrased as such, but they're quite rigid. (This is rather different
from the situation with, for example, pointers, where you are
explicitly not guaranteed that pointer types are
Well doing DP calculations at a GPU is a bit power and programmers
waste for now Geoff.
Remember, it's a 250 watt TDP monster that GPU, versus a Xeon 45 nm
chippie is 50 watt TDP.
How many units can do DP calculations? Like 40 or so.
According to my calculation that is practical a limit
Just for the sake of historicity for those less -- chronologically
enhanced-- than RGB or myself, the joke is attributed to Stroustroup, not
Thompson.
1. Ken Thompson wrote B (late 60's). Really minimal. Based largely on BCPL,
but minimal, ergo B.
2. Ken wrote Unix, in assembler.
3. Dennis
Kyle Spaans [EMAIL PROTECTED] writes:
Either way, I've recently started writing Conway's Game of Life in
C, as an exercise. I needed to figure out how to dynamically
allocate a 2D array. I found an answer on the comp.lang.c FAQ[1].
It's not terribly complex, but it seems to me like it's more
Perry E. Metzger wrote:
Robert G. Brown [EMAIL PROTECTED] writes:
On Tue, 26 Aug 2008, Michael H. Frese wrote:
C is not much better. I once worked a young computational
programmer for almost a week to get him to prove to himself that a C
source program couldn't walk through a 2-d
Ed Hill [EMAIL PROTECTED] writes:
And guarantees like the above make it rather easy for programmers to,
for instance, assemble their inputs in C++ and then, if they want,
call C, Fortran, assembly-optimized, or even GPU-/FPGA-implemented
routines to perform BLAS, (I)DFT, or other operations.
In message from Vincent Diepeveen [EMAIL PROTECTED] (Tue, 26 Aug 2008
00:30:30 +0200):
Hi Mikhail,
I'd say they're ok for black box 32 bits calculations that can do with
a GB or 2 RAM,
other than that they're just luxurious electric heating.
I also want to have simple blackbox, but 64-bit
Peter St. John [EMAIL PROTECTED] writes:
Just for the sake of historicity for those less -- chronologically
enhanced-- than RGB or myself, the joke is attributed to Stroustroup, not
Thompson.
1. Ken Thompson wrote B (late 60's). Really minimal. Based largely on BCPL,
but minimal, ergo B.
um, since array[i][j] is supposed to be the same as array (i * sizeof
(column declared length))[j], then I think yeah, the values should all be at
consecutive locations as Perry described. I just ran a similar experiment in
Visual Studio 2005, just what was on the box as I read this. I don't think
Perry,
Yes my mistake, // is in C99. Seems like yesterday :-)
And I didn't mean port so narrowly, didn't mean to imply he didn't write
an x86 kernel from scratch. And yeah my first x86 was System V, in the mid
80's.
Incidentally, while I understand that C with Classes preceeded C++, I use
the term
Dan Kidger wrote:
Gilad wrote:
It was proven by the same person who did the slides you
referred to,
when doing the same testing on IB DDR we got much better
results with
IB versus Quadrics. your theory does not really meet reality.
Care to describe to the list what these results
On 26 Aug 2008, at 12:53 pm, Perry E. Metzger wrote:
Perry E. Metzger [EMAIL PROTECTED] writes:
I have a copy of the C99 document and it is indeed required that the
locations be consecutive (though there can of course be padding for
alignment purposes if you have an array of structures).
If
On Tuesday 26 August 2008 16:38:11 Michael H. Frese wrote:
Vincent,
I have always said that C++ is computational science's cold fusion:
lots of power in, but no net gain.
C is not much better. I once worked a young computational programmer
for almost a week to get him to prove to himself
On Tue, 26 Aug 2008, Lux, James P wrote:
No. You are actually given guarantees about memory layout. They're not
phrased as such, but they're quite rigid. (This is rather different
from the situation with, for example, pointers, where you are
explicitly not guaranteed that pointer types are
Ed Hill [EMAIL PROTECTED] writes:
On Tue, 26 Aug 2008 14:29:11 -0400 Perry E. Metzger wrote:
I think part of the issue is that most people doing scientific
computing don't have computer science backgrounds, which is a
shame.
Formal CS training can certainly help but I don't view it as a
Robert G. Brown [EMAIL PROTECTED] writes:
I think that matrix allocation per se is contiguous because of how and
where it occurs at the compiler level. Immediate matrices allocated
like a[10][10] are just translated into displacements relative to the
stack pointer, are they not?
It depends.
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Alan Louis
Scheinine
Sent: Tuesday, August 26, 2008 2:27 PM
To: Ed Hill
Cc: Beowulf@beowulf.org
Subject: Re: [Beowulf] Stroustrup regarding multicore
I found some useful books in C++, I could not remember
Robert G. Brown wrote:
Any nontrivial program is
already almost by definition difficult to debug. Any nontrivial program
using multiple compilers or languages had damn well better have complete
separation at the shell level or debugging will go from difficult to
insanely difficult, really
Perry,
My first comment had nothing to do with whether C's 2-d arrays are
allocated contiguously in memory. I assume that they are, as was
implicit in my remark about doing the stepping by hand.
My comment has to do with the assembly code that the C compiler
generates to access an element
On Tue, 26 Aug 2008, Ed Hill wrote:
And I don't mean to sound like a cheezy motivational speaker but
if each of us is willing to study the texts a bit, read others' code,
and learn-by-doing then I don't think there is anything that prevents us
from becoming capable programmers. Or engineers.
Michael H. Frese [EMAIL PROTECTED] writes:
My first comment had nothing to do with whether C's 2-d arrays are
allocated contiguously in memory. I assume that they are, as was
implicit in my remark about doing the stepping by hand.
That was not what I got from your comment, but I'll take your
On Tue, 26 Aug 2008, Perry E. Metzger wrote:
Robert G. Brown [EMAIL PROTECTED] writes:
I think that matrix allocation per se is contiguous because of how and
where it occurs at the compiler level. Immediate matrices allocated
like a[10][10] are just translated into displacements relative to
David Mathog [EMAIL PROTECTED] writes:
Alan Louis Scheinine [EMAIL PROTECTED] wrote:
I forgot to mention, function calls in Fortran, C or C++ seem to kill
efficiency,
That's pretty much true for any language when those calls are inside
inner loops. There are exceptions though - inlined
Ed Hill wrote:
On Tue, 26 Aug 2008 14:29:11 -0400 Perry E. Metzger wrote:
I think part of the issue is that most people doing scientific
computing don't have computer science backgrounds, which is a
shame.
Formal CS training can certainly help but I don't view it as a
requirement. I've met
Robert G. Brown [EMAIL PROTECTED] writes:
On Tue, 26 Aug 2008, Perry E. Metzger wrote:
Perhaps, but don't most C programmers allocate such an array as a single
vector and then repack the indices?
I've never seen anyone allocate as a single vector and repack the
indices, though I'm sure
Robert G. Brown [EMAIL PROTECTED] writes:
On Tue, 26 Aug 2008, Perry E. Metzger wrote:
Robert G. Brown [EMAIL PROTECTED] writes:
I think that matrix allocation per se is contiguous because of how and
where it occurs at the compiler level. Immediate matrices allocated
like a[10][10] are just
Hello,
IMHO, it is better to call the BLAS or similiar libarary rather than programing
you own functions. And CUDA treats the GPU as a cluster, so .CU is not working
as our normal codes. If you have got to many matrix or vector computation, it
is better to use Brook+/CAL, which can show great
C++ books I have found useful
C++ Programming Language Third Edition, Bjarne Stroustrup
Advanced C++ Programming Styles and Idioms, James O. Coplien
Effective C++ Second Edition, Scott Meyers
C++ Coding Standards, Herb Sutter and Andrei Alexandrescu
Exceptional C++, Herb
35 matches
Mail list logo